Searching for Tuberculosis in the Pleural Space

Identifying tuberculous pleuritis is a common clinical problem with multiple pitfalls. We know that one third of patients with this condition can have a negative tuberculin skin test.¹ Pleural fluid culture results can be positive in < 25%.² Without treatment, the natural history of tuberculous pleuritis is spontaneous resolution with a high rate of recurrence (65%) of active tuberculosis disease.³ The AIDS epidemic has reminded us of the importance of identifying tuberculosis and treating it.⁴

The traditional answer to this clinical problem is to perform a needle pleural biopsy for both histologic study and culture, which can lead to the diagnosis of tuberculous pleuritis 86% of the time.⁵ These procedures, combined with cultures of pleural fluid and sputum, have been reported to provide microbiologic confirmation of Mycobacterium tuberculosis as often as 90% of the time.⁶ Tuberculosis culture results can take time, even with BACTEC (Becton Dickenson Microbiology Systems; Cockeysville, MD) and Gen-Probe (Gen-Probe; San Diego, CA) technology.⁷ A significant minority of patients (10 to 20%) will not have positive culture results or granulomas on biopsy specimen.

Four relatively new techniques have been reported to help make the diagnosis of tuberculous pleuritis: adenosine deaminase (ADA), lysozyme, interferon gamma, and polymerase chain reaction. Surprisingly, polymerase chain reaction has a relatively low sensitivity in pleural fluid (0.42 to 0.81)⁸⁹¹⁰ and is fairly expensive. The sensitivity of an elevated interferon level appears better (0.89 to 0.99),¹¹¹² but there have been relatively few studies of its use and the assay is expensive. Lysozyme pleural fluid to serum ratios have been reported to improve the sensitivity of the pleural fluid ADA.¹³ The test with the most data to support its use is the pleural fluid ADA.

ADA analysis is a simple and inexpensive colorimetric test that can be performed on body fluids.¹⁴ ADA, an enzyme that catalyzes the conversion of adenosine to inosine, is found in most cells. The isoenzyme ADA-2 is elevated significantly in pleural fluid with activated lymphocytes, such as from tuberculosis. False-positive results can occur with lymphoma, rheumatoid arthritis, systemic lupus erythematosus, and rarely adenocarcinoma.¹⁵ A different isoenzyme (ADA-1) is elevated in the presence of empyema.¹⁶ However, use of the isoenzyme assay is more expensive and not readily available. By not ordering an ADA assay of empyema fluid, one can avoid most false-positive results due to an elevated ADA-1.¹⁵

Several studies have suggested that an elevated pleural fluid ADA level predicts tuberculous pleuritis with a sensitivity of 90 to 100% and a specificity of 89 to 100% when the Giusti method is used.^111317181920 The reported cutoff value for ADA varies from 47 to 60 U/L. Choosing a lower value will increase sensitivity at the expense of specificity, as is elegantly discussed in the report by Riantawan and colleagues in this issue of CHEST (see page 97). This report is significant because it contains the largest group of HIV-positive patients with tuberculous pleural disease (37 total patients) reported in the literature. The ADA values in the HIV-positive patients did not differ significantly from those in the HIV-negative patients, contrary to what has been previously suggested by findings in a much smaller group of patients.¹²

The ADA measurement is used commonly in European and Asian countries where there is a higher incidence of tuberculosis. A group in Spain showed that patients with idiopathic pleural effusions could be followed without therapy for tuberculosis if the ADA level is < 43 U/L.²¹ There was no evidence of reactivation of tuberculosis in 10 years of follow-up despite 50% of the subjects having a positive purified protein derivative test.²¹ This strategy would be helpful in the United States, even with this country’s lower incidence of tuberculosis. The problem with using the ADA assay in a population with a lower incidence of tuberculosis is that the positive predictive value decreases; stated another way, there is a higher likelihood that a test would give a false-positive result. The downside of this could be unnecessary tuberculosis chemotherapy or a delay in the diagnosis of metastatic malignancy, neither of which is clearly a problem.

The other potential downside to the use of ADA measurement is that culture and drug sensitivity data would not be available, so patients with drug-resistant tuberculosis may be treated inadequately. As Riantawan and colleagues point out, the ADA assay should not serve as an alternate to biopsy and culture, but rather as an adjunct to allow earlier diagnosis and therapy. The assay would certainly be helpful in the 10 to 20% of patients who have a nondiagnostic workup despite biopsy.

Why is the ADA assay not used more in the United States? Perhaps the most common reason I hear from clinicians is that it is not readily available. This is perplexing since it is simple and inexpensive. Two laboratories (Fairview-University Laboratories, Minneapolis, MN, and Specialty Laboratories, Santa Monica, CA) offer an ADA assay that appears not to be the Giusti method, since the normal range in pleural exudates is reported as 1.6 to 9.2 U/L, which is much lower than the values reported in the literature. Therefore, there are no data with which to use this assay. To my knowledge, the only laboratory in the United States that performs the ADA assay commercially using the Giusti method is Micrim Laboratories, Fort Lauderdale, FL (K. Miller, PhD; personal communication; February 2, 1999). The laboratory’s personnel have been performing this assay for 5 years and report a range of 12 to 32 U/L in nonspecific pleural exudates. They suggest a cutoff value of 40 U/L, above which tuberculosis is highly likely. These values are consistent with the values reported in the literature. With a short, reliable turnaround time, a pleural fluid ADA value can be very helpful in the diagnosis of tuberculous pleuritis.